Project Summary The dissertation will investigate the main effects of 3 insurance benefit design features: 1. Financial requirements (e.g. copayments, coinsurance and deductibles), 2. Quantitative treatment limits (e.g. annual visit limits), and 3. Non-quantitative treatment limits (e.g. prior authorization), on specialty mental health utilization and expenditures. The study will also investigate how each benefit feature modifies the main effects of the other benefit features on mental health care utilization and expenditures. By providing a thorough and comprehensive investigation of how insurance affects specialty mental health utilization and expenditures, this dissertation will provide needed updates and expansions to the literature on this topic. It will also help policy makers, employers, and insurance companies better evaluate how policies that change insurance benefit generosity may also improve access to specialty mental health care. This study will link benefit design, claims, enrollment, provider supply and plan/employer characteristic data for commercial ?carve-in? self-insured plans from our community partner, Optum, one of the largest managed behavioral health organizations in the US. The data are from before (2009) and after (2011) implementation of the Mental Health Parity and Addiction Equity Act (MHPAEA) of 2008 (and publication of the Interim Final Rule in early 2010, requiring compliance starting January 2011). As a national mandate, MHPAEA contributes exogenous variation in benefit design. The study uses ?first-differences,? an analytic method that implicitly controls for patient characteristics that do not change over time (but that would otherwise bias the estimated effect of insurance). Using this approach, the study may avoid some selection bias, a common threat to determining the effects of insurance on utilization. Key outcome measures include counts of outpatient psychotherapy visits and inpatient days, as well as annual patient and plan expenditures for psychotherapy visits and inpatient days. Key explanatory variables include deductibles, copayments and coinsurance for psychotherapy visits and inpatient admissions, whether or not outpatient visit limits and inpatient day limits were required, and coinsurance and deductible penalties for failing to obtain prior authorization for outpatient or inpatient care when such authorization was required. Main models will be estimated using linear regression of change scores (2011-2009 values) with clustered standard errors controlled for by generalized estimating equations.